Download Left coronary pressure–flow relations of the beating and arrested

Cardiovascular Research 40 (1998) 88–95
Left coronary pressure–flow relations of the beating and arrested rabbit
heart at different ventricular volumes
Pieter Sipkema*, Johanna J.M. Takkenberg, Petra E.M. Zeeuwe, Nicolaas Westerhof
Laboratory for Physiology, Institute for Cardiovascular Research, Vrije Universiteit ( ICaR-VU), Amsterdam, The Netherlands
Received 16 July 1997; accepted 6 March 1998
Abstract
Objective: To study the effect of cardiac contraction on left coronary artery pressure–flow relations at different vascular volumes and
to compare these relations in the beating heart with those in the heart arrested in systole and diastole. Methods: Maximally vasodilated,
Tyrode perfused, rabbit hearts (n56) with an intra-ventricular balloon were used. The left coronary artery was separately perfused via a
cannula in the left main coronary artery. The slopes and the intercepts of left coronary pressure-flow relations were determined in the
beating and arrested heart at different chamber volumes. A 3-factor design with repeated measures was used to compare the effect of three
factors: phase of contraction (systole and diastole), chamber volume (V0 and V1 , left ventricular end-diastolic pressure 1.4 and 20 mm Hg,
respectively) and the type of contraction (beating and arrested; a measure of capacitive effects). Results: The phase of contraction has a
significant effect on the intercepts (.40 mmHg, p50.00032) but not on the slopes of the pressure–flow relations. Chamber volume had a
small effect on the intercepts (,5 mm Hg, p50.037), but not on the slopes of the pressure–flow relations. The type of contraction has a
significant effect on the slopes (|10%, p50.00021) but not on the intercepts of the pressure–flow relations. Conclusions: In the isolated
Tyrode perfused rabbit heart left coronary pressure–flow relations are mainly determined by contraction, while left ventricular chamber
volume and capacitive effects contribute little.  1998 Elsevier Science B.V. All rights reserved.
Keywords: Diastolic arrest; KCl; Barium contracture; Isolated heart; Left ventricular pressure; Rabbit
1. Introduction
Coronary arterial inflow is impeded in systole as a result
of cardiac muscle contraction. Previously, we have hypothesized that this diminished flow is the result of the varying
mechanical properties of the cardiac muscle during the
heart cycle [1–6]. In the ventricular cavity, contraction
results in a time varying pressure-chamber volume relation, i.e., the time varying elastance [7,8]. In the coronary
vessels, we assume cardiac contraction to have the same
effect, resulting in a time varying pressure–vascular volume or pressure–vascular cross-sectional area relation [9].
The decrease in cross-sectional area of the coronary
vessels during systole results in an increase in resistance,
and in a ‘pumping action’ or capacitive effect [1]. Left
*Corresponding author. Tel.: 131 (20) 444 8110; Fax: 131 (20) 444
8255; E-mail: [email protected]
ventricular pressure, as used by the waterfall model [10]
and the intra myocardial pump model [11–13], play a
minor role in our concept.
Earlier, we studied the effect of cardiac contraction on
coronary arterial inflow in the maximally dilated [5] and
autoregulating bed [4], while keeping constant arterial and
venous pressures. In these studies, variation of left ventricular pressure by changing left ventricular chamber
volume had little effect on the impeding of coronary flow.
These findings are in line with the varying elastance
concept. However, only one single perfusion pressure was
used. Coronary pressure–flow relations in diastole were
determined in several studies [14–16]. However, to the
best of our knowledge, no studies have been performed
where coronary pressure–flow relations were measured
simultaneously both in systole and diastole both in the
Time for primary review 35 days
0008-6363 / 98 / $ – see front matter  1998 Elsevier Science B.V. All rights reserved.
PII: S0008-6363( 98 )00119-9
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
beating and in the arrested heart at different volumes. The
changes in vessel diameters during the cardiac contraction
result not only in vascular resistance changes but also in
vascular volume changes (capacitive effects) and tend to
decrease arterial inflow mainly in early contraction [1,17].
These capacitive effects can be derived by comparing
diastolic and systolic left coronary pressure–flow relations
in the beating with those obtained in the arrested heart.
In the present study we test the hypothesis that tissue
stiffness is the main determinant of coronary flow. The
stiffness change due to contraction is much larger than the
change in stiffness resulting from volume changes. Therefore contraction (i.e., the difference between diastolic and
systolic muscle) is predicted to have a larger effect on flow
than stretch (i.e., an increase in ventricle volume).
Altogether three factors that determine coronary pressure–flow relations were studied: phase of contraction
(systole and diastole), left ventricular chamber volume (V0
and V1 ) and type of contraction (beating and arrested).
2. Methods
The investigation conforms with the Guide for the care
and use of laboratory animals published by the US
National Institutes of Health (NIH Publication No. 85-23,
revised in 1985) and is performed according to the
guidelines of the D.E.C. (Animal Experiments Committee
of the Free University of Amsterdam, The Netherlands).
2.1. Heart preparation
Male Flemish Giant rabbits (n56), weighing 5.9260.35
kg (mean6sem), were anaesthetized with Fluanisone (9
mg / kg) and fentanyl citrate (0.3 mg / kg) (Hypnorm,
Janssen Pharmaceutica B.V., Tilburg, The Netherlands),
injected intramuscularly. Subsequently, sodium pentobarbital (30 mg) with benzyl alcohol (4.5 mg, Nembutal, Sanofi
B.V., Maassluis, The Netherlands) was injected in an ear
vein and the heart was exposed. Heparin (2500IU, Leo
Pharmaceutical Products B.V., Weesp, The Netherlands)
was administered to avoid blood coagulation. The aorta
was cannulated, the heart was excised quickly and the
aorta cannula was connected to the set up. The perfusion of
the heart was discontinued only for a few seconds.
The heart was perfused (not recirculated) with a Tyrode
solution (128.3 mM NaCl, 4.7 mM KCl, 1.36 mM
CaCl 2 .2H 2 O, 1.05 mM MgCl 2 .6H 2 O, 20.2 mM NaHCO 3 ,
0.42 mM NaH 2 PO 4 .2H 2 O, 11.1 mM C 6 H 12 O 6 .H 2 O) containing adenosine (10 mM) to maximally dilate the coronary vascular bed, in order to establish constant vasomotor tone. The perfusate was bubbled with a mixture of
95% O 2 and 5% CO 2 at 288C and pH 7.3–7.4. Before
entering the heart the Tyrode solution was passed through
a 3 mm filter (Millipore S.A., Molsheim, France). To drain
89
the Thebesian flow a cannula was introduced through the
apex of the left ventricle.
The atrio-ventricular node was destroyed by crushing
the nodal tissue. The heart was paced by two electrodes,
one attached to the apex cannula and the other to the right
ventricle. Pace frequency was set at approximately 0.7 Hz,
just above the ventricular rhythm. A latex balloon, filled
with distilled water was placed in the left ventricle via the
left auricle and through the mitral valve. The heart was
beating isovolumically. Left ventricular balloon pressure
(LVP) was measured using a pressure transducer (Gould
P23ID, USA).
2.2. The modified Langendorff set up ( Fig. 1)
The Langendorff set up consisted of two separate
perfusion lines to change the flow to the left heart
specifically. The two perfusion lines were connected to
two Tyrode-filled reservoirs (a and b in Fig. 1) to obtain
different perfusion pressures (see protocol). A rapid
pneumatic switch was used to alternate perfusion of the
heart from either reservoir (a) or (b). Both the function
generator and the pneumatic switch were controlled by a
digitimer. The right perfusion line, containing a thermistor
and an electromagnetic flow probe (EM, Skalar Instruments, The Netherlands), terminated in the aortic cannula.
The left coronary perfusion line contained a transit-time
flow probe (TT, Transonic Systems, Ithaca, New York,
U.S.A.) and was connected to a small cannula (see inset
Fig. 1), that was inserted through the aortic cannula into
the left main stem, where it was wedged in position and
secured with a suture. Left coronary perfusion pressure
(Pc, Gould P23ID USA) was measured at the proximal
side of the left main stem cannula.
To check whether the left main stem cannula was tight
in place in the left main stem, two tests were performed.
Ink was injected into the heart through the left or right
perfusion line respectively. Fluid leakage was found to be
unlikely when the section of the heart not perfused with
ink did not color. Second, zero flow pressures were
measured in the left main stem by occluding both perfusion lines and then occluding only the left perfusion line. If
the difference between these two zero-flow pressures was
within 5 mmHg we assumed that perfusion of the left
coronary bed occurred solely through the left main stem
cannula. When leakage was found, the cannula was repositioned and the three tests were repeated.
2.3. Protocol ( Fig. 2)
The experiments were divided in four sections (see
bottom section of Fig. 2). Each section is divided in three
volume settings and at each volume a left coronary
pressure flow relation is determined. In the first section the
heart was perfused with Tyrode and beating. In the second
90
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
Fig. 1. Experimental set-up of the isolated heart. The left coronary artery cannula is shown in the inset. The pneumatic switch, controlled by the digitimer,
connects both perfusion lines to either reservoir a (reference pressure) or reservoir b (variable pressure). PID unit5Proportional Integral Derivative control
unit; flow (TT)5transit time flow meter; flow (EM)5electromagnetic flow meter; Pm5pressure transducer left coronary artery.
section the heart was arrested in diastole (static diastole) by
changing the perfusion solution to a Tyrode solution
containing 20.0 mM KCl (NaCl concentration was lowered
to 113 mM). In the third section the heart was perfused
with normal Tyrode and beating again to check the
stability of the preparation. Pressure–flow relations had to
be statistically not significantly different from the first
section of the experiment, in order to accept the experiment. In the fourth and final section of the experiments, the
heart was stopped in systole (static systole) by lowering the
calcium concentration of the Tyrode (0.078mM CaCl 2 .2H 2
O) and adding 2.0–3.0 mM BaCl 2 (Barium contracture).
Since BaCl 2 is known to cause vasoconstriction, adenosine
(50 mM) and Isoproterenol (55 nM) (Sigma I-2760, St.
Louis MO, USA) were also added. In pilot experiments we
found that left ventricular pressure in static systole was
approximately 80% of the peak systolic left ventricular
pressure in the beating heart. Systolic left ventricular
pressure in the beating heart (first and third section of the
experiments) was therefore lowered to 80% of the original
value by adding Verapamil (Sigma V-4629, St. Louis MO,
USA).
In each section ventricular chamber volume was
changed (V0 , V1 , V0 ). The first and last chamber volume
settings were identical to test the stability of the heart.
At each volume setting a perfusion pressure protocol
consisting of nine steps (lasting 6 s each, sufficient to
reach a steady state, Fig. 2 top part) was generated for the
left coronary bed to determine a pressure–flow relation.
Perfusion pressure was after each step returned to the
reference pressure (90 mmHg) to check the stability of the
preparation by testing if the flow rate at the reference level
was constant. When the reference flow rate was not stable
($10% difference compared to the starting flow rate), the
preceding and following steps of the protocol were not
used for analysis.
At the end of the experiment the wet and dry (608C, 48
h) were measured. By calculating the dry–wet weight ratio,
formation of edema could be quantified.
2.4. Measurements and analysis of data
Left ventricular pressure (mmHg), left coronary pressure
(Pc, mmHg), and flow rate (ml / min) were recorded on a
Thermal Array recorder WR3600 (Graphtec Corporation,
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
91
Fig. 2. Experimental protocol. The entire protocol is indicated at the bottom. Each condition is divided in three parts: V0 , V1 and V0 . The dashed lines
indicate a blow-up of the first section of the protocol line. A nine step pressure protocol was generated at different ventricular volumes in both the beating
and the arrested (systolic and diastolic) heart. Steps in perfusion pressure are applied between 40 and 130 mm Hg. Between steps pressure is at its reference
level of 90 mm Hg. The flow values measured at the reference level are used to check the stability of the preparation. The return to beating is also used for
stability testing. From each step-protocol a pressure–flow relation is determined.
Tokyo, Japan). Left coronary pressure and flow rate values
were also stored (sampling rate 200 Hz) on a personal
computer (M290S, Olivetti). For each flow rate, the
pressure drop across the left main stem cannula was
calculated and subtracted from the measured left coronary
perfusion pressure (Pm) in order to determine the true left
coronary pressure (Pc) at the tip of the cannula.
Left coronary pressure (Pc)–flow relations were made
by applying linear regressions (Left coronary pressure is
the independent variable). The Akaike information criterion [18] was employed to investigate if relations could be
fitted better by a quadratic equation than a linear one. The
intercepts and slopes of the linear regression analysis were
used in further statistical comparisons.
Left coronary pressure–flow relations in the beating and
the arrested hearts were analyzed at selected left ventricular chamber volumes that resulted in end-diastolic
pressures of 1–2 mmHg (V0 ) and 20 mmHg (V1 ) and in
systolic left ventricular pressures of approximately 80
mmHg (V0 ) and 115 mmHg (V1 ).
The effects of cardiac contraction, ventricular chamber
volume and type of contraction on slopes and intercepts of
left coronary pressure–flow relations were investigated
using analysis of variance with repeated measures, i.e.,
each heart is investigated under the three conditions. The
analysis extracts the variance for the three main effects:
(A) phase of contraction, (B) chamber volume and (C)
type of contraction. (Systat software, release 4, Evanston
IL 60201). Mean values of slopes and intercepts are given
with Standard Error. Significance was chosen at p,0.05.
3. Results
Six successful experiments on isolated rabbit hearts
were performed. The left main stem cannula provided the
perfusion fluid to the left heart and did not leak. Ink
injected in the left main stem cannula only showed up in
the left heart. Ink injected in the perfusion line (right
perfusion) did not become visible in the left heart. The
mean measured zero flow intercept (occlusion of the left
heart perfusion flow) in the beating heart was 0.3360.5
mmHg. We analyzed 24 step-protocols each consisting of
nine steps in the six experiments. We left out 15 steps (7
steps in experiment 6) in the analysis because the control
flow before and after the step was different.
Two left ventricular chamber volumes were studied: V0
and V1 . In the beating and diastolic arrested hearts V0
resulted in left ventricular end-diastolic pressures of
1.060.9 mmHg and 1.860.5 mmHg respectively, while at
V1 these values were 19.861.5 mmHg and 20.361.3
mmHg, respectively. In the systolic arrested hearts we
measured left ventricular pressures of 78.065.2 mmHg at
V0 and 111.764.9 mmHg at V1 , pressures not significantly
different from the systolic left ventricular pressures in the
beating hearts (78.263.2 mmHg and 119.7613.2 mmHg,
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
92
respectively). Wet weight of the hearts (n56) was
14.561.0 gram and dry weight was 2.6360.2 gram. The
dry–wet weight ratio was 0.1860.01.
Table 1 shows slopes and intercepts (individual and
their mean values) of left coronary pressure–flow relations
in diastole and systole at the two ventricular chamber
volumes (diastolic pressure at V0 was 1–2 mmHg and at V1
20 mmHg) in the beating and arrested heart for all
experiments. The p-values from analysis of variance are
given at the bottom.
Fig. 3 and Fig. 4 show systolic and diastolic pressure–
flow relations of the left main coronary artery in the six
beating and arrested hearts, respectively, at two chamber
volumes. Differences in pressure–flow relations between
the two chamber volumes are small. The intercept of the
pressure–flow relations was found to depend on the
contraction phase (diastole / systole, p50.00032) and
slightly on the chamber volume of the ventricle ( p5
0.037). Post hoc tests on intercepts show that all individual
comparisons between ’diastole and systole’ are significant.
Individual comparisons of intercepts between V0 and V1 are
not significant in systole, but are significant in diastole.
The pressure–flow relations of beating and arrested
hearts at the smallest chamber volume (V0 ) are compared
in Fig. 5. Diastolic flow (indicated with D) is smaller at the
same perfusion pressure in the arrested hearts (open
square) compared to the beating heart (filled square).
Systolic flow (indicated with S) in the arrested hearts
(open squares) and the beating hearts (filled squares) are
about the same. Only the slopes of pressure–flow relations
comparing the beating and arrested heart were slightly, but
significantly different, a small clockwise rotation was
found. The intercepts were not different ( p50.43).
Normalized pressure–flow relations at chamber volume
V0 of all six experiments are displayed in Fig. 6. It is
evident that in both the beating and the arrested hearts
Fig. 3. Left coronary pressure–flow relations at left ventricular volumes
V0 and V1 in the beating hearts (n56). Each panel contains four relations.
The filled symbols are relations at V0 and the open symbols at V1 . The two
upper relations are diastolic and the two lower relations are systolic. Pc
(mmHg)5left coronary artery pressure; Flow5left coronary artery flow
rate. The roman numbers are the experiment numbers.
systolic flow is significantly impaired compared to diastolic flow, at all perfusion pressures studied.
4. Discussion
The results from this study show that left coronary
pressure–flow relations in the maximally dilated left
coronary bed of a Tyrode-perfused rabbit heart differ
significantly between systole and diastole, i.e., contraction
causes a parallel shift to lower flows. Increase in left
ventricular chamber volume, corresponding with a diastolic ventricular pressure rise from |1 to |20 mmHg, did
result in a small but significant decrease in the flow rate.
Table 1
(A) Slopes (ml / min / mmHg) and pressure axis intrcepts (mm Hg) of left coronary pressure-flow relations (Flow5(Pc -Pi )*slope) in beating and arrested
hearts at two different volumes (V0 , V1 );(B) p-values of slopes and intercepts of left coronary pressure–flow relations
(A)
Beating (systole)
Slope
Beating (diastole)
Arrested (systole)
Arrested (diastole)
Intercept
Slope
Slope
Intercept
Slope
Intercept
Intercept
Experiment [
V0
V1
0
V1
V0
V1
V0
V1
V0
V1
V0
V1
V0
V1
V0
V1
I
II
III
IV
V
VI
Mean
SEM
1.42
0.87
1.19
0.86
0.89
0.94
1.02
0.09
1.46
0.94
1.25
0.77
0.97
1.04
1.07
0.10
49.6
35.6
36.9
42.0
41.7
39.1
40.8
2.0
51.7
43.1
38.6
37.1
43.7
41.6
42.7
2.1
1.44
1.86
1.24
0.98
0.87
0.92
1.21
0.16
1.43
1.75
1.26
0.90
0.89
1.00
1.21
0.14
-14.9
-4.7
-17.3
-8.3
-32.6
-23.3
-19.2
5.3
-14.2
-1.3
-13.4
-9.2
-28.0
-19.7
-14.3
3.7
1.10
0.94
0.98
0.67
0.76
0.75
0.87
0.07
1.10
0.95
0.91
0.63
0.72
0.76
0.85
0.07
35.6
38.0
32.7
42.4
37.2
46.5
38.7
2.1
40.4
41.1
37.3
43.3
41.7
44.5
41.4
1.0
1.42
1.50
1.16
0.86
0.83
0.98
1.13
0.12
1.33
1.52
1.16
0.84
0.83
0.97
1.11
0.11
-3.4
3.6
-6.1
1.4
-15.3
-10.9
-5.1
2.9
-5.3
6.4
-3.7
4.5
-11.3
-4.7
-2.4
2.7
(B)
P-values
Diastole / systole
Volume V0 /V1
Beating / arrested
Slope
Intercept
0.10
0.00032
0.89
0.037
0.00021
0.43
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
Fig. 4. Left coronary pressure–flow relations at V0 and V1 in the hearts
(n56) both arrested in systole and in diastole. Pc (mmHg)5left coronary
artery pressure; Flow5left coronary artery flow.
Furthermore, comparing beating with arrested hearts,
shows that the slopes of the left coronary pressure–flow
relations in the arrested heart are smaller than in the
beating heart (clockwise rotation).
4.1. Discussion of methods
The mean dry to wet ratio in this study was found to be
0.1860.01. In other studies in our laboratory [1,19] a
similar value (0.1760.05) was found. The steady conditions of the hearts were checked by measuring the flow
throughout the experiment at the reference pressure. No
significant change was found. Any increase in edema
throughout the experiment would be reflected in a decrease
in this reference flow [16].
Fig. 5. Comparison of left coronary pressure–flow relations in the beating
versus the arrested hearts (n56) at the smallest left ventricular volume
(V0 ). Pc5left coronary artery pressure; Flow5left coronary artery flow
rate.
93
Fig. 6. Normalized pressure–flow relations of all experiments. On the left,
results of measurements in the beating heart are displayed. On the right,
results of measurements in the arrested heart are displayed. ‘% of Flow at
Pc90’ is defined as measured left coronary artery flow / reference diastolic
left coronary artery flow at a left coronary artery pressure of 90
mmHg)*100%; Pc5left coronary artery pressure.
In the beating hearts we added adenosine to the perfusion medium in order to maximally dilate the coronary
bed, i.e., flow mediated dilation and myogenic effects
could not influence the pressure–flow relationships. In the
diastolic arrested heart there is some remaining tone due to
the fact that we did not counteract the potassium constriction completely and a small contribution of the myogenic
response or shear stress to the pressure–flow relation could
be possible. This might explain the slightly smaller flows
in the diastolic arrested hearts as compared to diastolic
flows in the beating hearts (Fig. 5).
We used Barium contracture to study mechanical interaction in the systolic arrested heart. It could be reasoned
that the tissue deteriorates during this intervention. However, the metabolic and functional consequences of the
Barium induced contracture are studied by Shibata et al.
[20]. They used Tyrode perfused rabbit hearts at 378C and
found that after Barium contracture the tissue is stable for
approximately 30 min with a perfusion flow of 35 ml / min.
We used a lower temperature (278C) during our experiments. The Q 10 for heart metabolism is 2.1 [21] and
therefore a flow of |17 ml / min (35 divided by 2.1) would
have been sufficient. Our systolic arrest protocol lasted less
than 30 min. and our perfusion flow during systolic arrest
at reference pressure was minimally 31.9 ml / min. Thus,
conditions were chosen to prevent cardiac deterioration.
Temporarily, during our step protocol, we went down to
lower flows, but the duration (six sec) was less than the
actual delay time for the start of autoregulation. We also
tested cardiac conditions in the step protocol (Fig. 2) and
had stable pressure development. We conclude that deterioration of the tissue during the period of Barium arrest is
unlikely.
We applied the Akaike criterion to all experiments to
investigate whether the pressure–flow relations were best
described by a linear or quadratic function. This was done
because it seemed that the pressure–flow relations in Figs.
94
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
3 and 4, tended to be curvilinear, as suggested in the
literature [15,22–24]. Curvilinearity is accentuated at
lower pressures and may reflect changes in the number of
perfused channels as well as the size of the individual
channels. Hanley et al. [23] support the latter option.
Curvilinear systolic coronary pressure–flow relations were
predicted by a model for the effect of cardiac contraction
on coronary blood vessels [9] developed in our institute.
However, although in 70% of the cases a curvilinear
relationship fitted the data somewhat better according to
the Akaike criterion over the pressure range studied, it can
be seen that the difference between a linear and a quadratic
relation was very small. Therefore, and also for the sake of
simplicity, we decided to use linear regression for the
comparison of different left coronary pressure–flow relations. The intercept of the linearized pressure–flow relation
is then not necessarily the zero flow pressure intercept.
4.2. Diastolic /systolic differences in beating and
arrested hearts
We found that cardiac contraction causes a significant
shift in left coronary pressure–flow relations to lower
flows both in the beating and in the arrested heart.
Livingston and colleagues [24] observed similar shifts of
the coronary pressure–flow relations in the isolated maximally dilated perfused septum of the dog heart comparing
the diastolic (not beating) and tetanized states. Using linear
fits they found that the intercept between tetanized and
diastolic relations was significantly different while the
slope was not. In a more recent study in the same
experimental model by Livingston et al. [25], a small
portion of this rotation was found to be attributable to the
Gregg effect (The increase of coronary perfusion causes an
increase in contractility and myocardial oxygen consumption [26]). The pressure–flow relation was therefore not
affected uniformly, because at higher perfusion pressure,
and thus higher contractility, flow is relatively more
affected so that the pressure–flow relation rotates clockwise. We analyzed the Gregg effect, as the relation
between systolic left ventricular pressure and perfusion
pressure(Ps /Pp ), in the beating and arrested heart at V0 . We
found that the mean values of the slope of the relations to
be 0.073760.282 and 0.18660.114 for beating and arrested conditions resp. (NS, p50.313). Thus, the Gregg
phenomenon does not explain the relative rotation between
the beating and arrested conditions in systole in our
experiments.
The effect of capacitive flow on diastolic left coronary
pressure–flow relations is expected to be negligible, because we used a very low heart rate and we measured
diastolic flow at the end of diastole, and at this point in
time most capacitive effects, if any, have already disappeared.
4.3. The arrested heart
Potassium chloride (used for diastolic arrest) causes
vasoconstriction. In a separate study (not published) we
investigated the effect of potassium and adenosine on the
diameter of isolated rabbit left coronary arteries (n57) at
288C. The diameter was 86.863.0% (from the maximal
diameter) during perfusion with 20 mM potassium and
after adding 10 mM adenosine 93.363.2%. Since we used
these concentrations of potassium and adenosine in the
present study, we cannot exclude some remaining vasoconstriction. The significant reduction of the intercept of the
diastolic pressure–flow relation as observed in the arrested
hearts compared to the beating hearts (reduction of diastolic flow by approximately 20% at reference flow) can
thus be explained by the incomplete vasodilatation, since a
diameter reduction of 6% will result, on basis of Poiseuille’s law, in a flow reduction of approximately 22%.
Barium is a good model for systolic arrest [27,28], but
known to constrict coronary arteries [29]. In pilot studies
we studied Barium-induced contraction of rabbit coronary
arteries, and found that effective dilatation was established
with a mixture of adenosine (50 mM) and isoproterenol
(55 nM) at 288C [29]. We used that same mixture in the
present study and therefore concluded that there was no
barium-induced constrictive effect on the diameter of
coronary vessels. This mixture of adenosine and isoproterenol was also used by Goto et al. [30], but their
study was performed at 378C. At that temperature we
found in the same pilot data that this mixture was not
sufficient to counteract the Barium effect on the vessels.
4.4. Effect of chamber volume
The effect of chamber volume did overall result in a
small but significant decrease in flow rate (intercept p5
0.037, slope NS). In the arrested hearts in diastole and
systole the effect of chamber volume was not significant
( p50.16). We used a left ventricular pressure of 80 mmHg
and 115 mmHg. Iwanaga et al. [30] discussed the role of
the subepicardium in extramural coronary flow pulsatility.
The subepicardium may shunt in systole most of the
retrograde flow from the subendocardium. Thus the observation that capacitive effects in global flow are small, as
observed in the present study, may be due to the shunting
phenomenon as discussed by Iwanaga et al. [30].
In summary, in this experimental model left coronary
pressure–flow relations are significantly different in systole
and diastole, left ventricular chamber volume has little
influence on these relations, and capacitive effects are
small. Ideally one could think that the in vivo situation is
best suited to study the sensitivity of coronary pressure–
flow relations to interventions, but this is not always the
case when one wants to study certain phenomena without
P. Sipkema et al. / Cardiovascular Research 40 (1998) 88 – 95
the confounding contributions of other factors. Therefore,
the conditions were chosen to be able to test the hypothesis. The results of this study may, for instance, be applied
to the stenotic coronary circulation where the bed distal to
the stenosis is (maximally) dilated. Since we found that
increased contractility has a stronger impeding effect on
flow than an increase in ventricular volume, an increase in
stroke volume should be obtained by a volume increase
rather than by increased contractility. Future studies in this
model will be directed to determine these effects on left
coronary pressure–flow relations in different layers of the
myocardium.
Acknowledgements
This study was supported in part by the Netherlands
Organization for Scientific Research (NWO GB-MW grant
[ 902-16-175).
References
[1] Bouma P, Sipkema P, Westerhof N. Coronary arterial inflow
impediment during systole is little affected by capacitive effects. Am
J Physiol 1993;264:H715–H721.
[2] Bouma P, Sipkema P, Westerhof N. Vasomotor tone affects diastolic
coronary flow and flow impediment by cardiac contraction similarly.
Am J Physiol 1994;266:H1944–H1950.
[3] Krams R, van Haelst ACTA, Sipkema P, et al. Can coronary
systolic–diastolic flow differences be related to left ventricular
pressure? A study in the isolated rabbit heart. Basic Res Cardiol
1989;84:149–159.
[4] Krams R, Sipkema P, Westerhof N. The varying elastance concept
may explain coronary systolic flow impediment. Am J Physiol
1989;257:H1471–H1479.
[5] Krams R, Sipkema P, Westerhof N. Contractility is the main
determinant of coronary systolic flow impediment. Am J Physiol
1989;257:H1936–H1944.
[6] Krams R, Sipkema P, Westerhof N. Coronary oscillatory flow
amplitude is more affected by perfusion pressure than ventricular
pressure. Am J Physiol 1990;258:H1889–H1898.
[7] Suga H, Sagawa K, Shoukas AA. Load independence of the
instantaneous pressure–volume ratio of the canine left ventricle, and
effects of epinephrine and heart rate on the ratio. Circ Res
1973;49:314–322.
[8] Suga H, Sagawa K. Instantaneous pressure–volume relationships
and their ratio in the excised, supported canine left ventricle. Circ
Res 1974;35:117–126.
[9] Vis MA, Sipkema P, Westerhof N. Modelling pressure–flow relations of cardiac muscle in diastole and systole. Am J Physiol
1997;272:H1516–H1526.
[10] Downey JM, Kirk ES. Inhibition of coronary flow by vascular
waterfall mechanism. Circ Res 1975;36:753–763.
95
[11] Arts MGJ, Reneman RS. Interaction between intra myocardial
pressure (IMP) and myocardial circulation. J Biomech Eng
1985;107:51–56.
[12] Spaan JAE, Breuls NPW, Laird JD. Forward coronary flow normally
seen in systole is the result of both forward and concealed back
flow. Basic Res Cardiol 1981;76:582–586.
[13] Spaan JAE, Breuls NPW, Laird JD. Diastolic–systolic coronary flow
differences are caused by the intra myocardial pump action in the
anesthesized dog. Circ Res 1981;49:584–593.
[14] Bellamy RF. Diastolic coronary artery pressure–flow relations in the
dog. Circ Res 1978;43:92–101.
[15] Aversano T, Klocke FJ, Mates RE, et al. Preload induced alterations
in capacitance free diastolic pressure–flow relationship. Am J
Physiol 1984;246:H410–H417.
[16] Van Dijk LC, Krams R, Sipkema P, et al. Changes in coronary
pressure–flow relation after transition from blood to Tyrode perfusion. Am J Physiol 1988;255:H476–H482.
[17] Gregg DE, Green HD. Registration and interpretation of normal
phasic inflow into a left coronary artery by an improved differential
manometric method. Am J Physiol 1940;130:114–125.
[18] Landaw EM, DiStefano JJ. Multiexponential, multicompartmental
and noncompartmental modelling. II. Data analysis and statistical
considerations. Am J Physiol 1984;246:R665–R677.
[19] Hak JB, van Beek JHGM, van Wijhe MH, et al. Influence of
temperature on the response time of mitichondrial oxygen consumption in the isolated rabbit heart. J Physiol 1992;447:17–31.
[20] Shibata T, Berman MR, Hunter WC, et al. Metabolic and functional
consequences of barium-induced contracture in rabbit myocardium.
Am J Physiol 1990;259:H1566–H1574.
[21] Hak JB, van Beek JH, van Wijhe MH, et al. Influence of temperature on the response time of mitochondrial oxygen consumption in
isolated rabbit heart. J Physiol 1992;447:17–31.
[22] Klocke FJ, Mates RE, Canty Jr. JM, et al. Coronary pressure–flow
relationships. Controversial issues and probable implications. Circ
Res 1985;56:310–323.
[23] Hanley FL, Messina MR, Grattan MT, et al. The effect of coronary
inflow pressure on coronary vascular resistance in isolated dog heart.
Circ Res 1984;54:760–772.
[24] Livingston JZ, Resar JR, Yin FCP. Effect of tetanic myocardial
contraction on coronary pressure–flow relationships. Am J Physiol
1993;265:H1215–H1226.
[25] Livingston JZ, Halperin HR, Yin FCP. Accounting for the Gregg
effect in tetanized coronary arterial pressure–flow relationships.
Cardiovasc Res 1994;28:228–234.
[26] Feigl EO. Coronary physiology . Phys Rev 1983;63:13–18.
[27] Judd RM, Levy BI. Effects of barium-induced cardiac contraction
on large- and small-vessel intra myocardial blood volume. Circ Res
1991;68:217–225.
[28] Munch DF, Comer HT, Downey JM. Barium contracture: a model
for systole. Am J Physiol 1980;239:H438–H442.
[29] Sipkema P, van der Linden PJW, Zeeuwe PEM, et al. Contraction
and dilatation of the rabbit coronary artery due to BaCl 2 at different
temperatures. J Vasc Res 1994;31(S1):47.
[30] Iwanaga S, Ewing SG, Hussini WK, et al. Changes in contractility
and afterload have only slight effects on subendocardial systolic
flow impediment. Am J Physiol 1995;269:H1202–1212.